Tell me i didnt waste my money

[allieboy3]

I did not overclock

I was running 2 9800gt in sli and upgraded to msi 6850. There was no difference in my 3dmark 06 benchmark. both at 17,000

I have a i7 - 920 6 gigs of ram.

please tell me its because im now running dx11. Did i see much of a difference in bad co2 to say wow... nope

thanks

 

[Mistwalker]

That does seem low, this site seemed to get about 20,000 depending on the resolution, as did this one.

Make sure you run Driver Sweeper to clean out any old NVidia and Physx files, and that the 6850 is running at x16 PCI speed (use GPU-Z) would be my first suggestions.

 

[VirtualLarry]

That does seem low. My Q6600 @ 3.6, with one HD4850, gets around 12,000, and with two HD4850s, gets 19,500 in 3dMark06.

Another rig that I recently build, with a Q9550 @ 3.5, and two HD6870s in CF, gets a little under 20000.

 

[vshin]

Don't run programs like MSI Afterburner when you bench on 3DMark because it lowers your score for some reason.

 

[crisium]

I'm not sure about the validity of an 11,000 score, but a 6850 is probably only a slightly faster than 9800GT SLI. I wasn't able to find a direct comparison, but if we accept that a 9800GX2 is a bit faster than 9800GT SLI and a 5850 is a bit faster than a 6850, then we can use this:

http://tinyurl.com/239e9jj

Not the exact comparison you're looking for, but I'd imagine similar results.

3DMark is at 1280x1024 correct? If so, then the advantage of moving from 512mb to 1GB of ram probably isn't there. 3DMark also scales well with SLI. So from the benchmark I linked we can see that sometimes the 5850 (6850) is decisively faster, and other times it's the same speed as the 9800GX2 (9800GT SLI). Why can't 3DMark be one of the times they are the same speed?

Not sure about 11,000 though.

 

[notty22]

I agree with your gpu power being about the same from 9800 sli to 6850.
But that 11,000 is not correct.
I could score between 13000-17000 with a o/c 4770, depending on both o/c's,
Do the whole uninstall driver sweeper , re-install. Believe the best drivers are 10.10e for 6850.
After that is done, proven to have helped. o/c a 100mhz

 

[allieboy3]

please forgive i have corrected my original post. clocked around 17K
thanks for replys

 

[TemjinGold]

You did not waste your money.

 

[WelshBloke]

Don't run programs like MSI Afterburner when you bench on 3DMark because it lowers your score for some reason.

Really? Thats weird, I use Afterburner for automatic 2d/3d clocks, but I've not run 3DMark for a long time.

 

[tyl998]

I heard that 3dMark06 is more CPU-bottlenecked than GPU-bottlenecked.

 

[dust]

That does seem low. My Q6600 @ 3.6, with one HD4850, gets around 12,000, and with two HD4850s, gets 19,500 in 3dMark06.

Another rig that I recently build, with a Q9550 @ 3.5, and two HD6870s in CF, gets a little under 20000.

That seems low for 6870 cf, what resolution were you benching at?

 

[Rhoxed]

That seems low for 6870 cf, what resolution were you benching at?

It seems spot-on with his CPU, my 940BE @ 3.8 and 3 4850's score 21000, while my 1090T @ 4.2 with the same 3 4850's scores scores 24255 - 3dmark 06 is all about cpu speed now

 

[RussianSensation]

please tell me its because im now running dx11. Did i see much of a difference in bad co2 to say wow... nope

thanks

Outside of tessellation, there are really not many games where DX11 by itself brings benefits (I can think of Civilization 5 and BattleForge where the cards run faster in DX11). Most likely, we won't see "pretty" DX11 games until next generation of consoles (so 2013+).

 

[digitaldurandal]

You should compare to the actual in game performance and features you get and not to 3dmark. If you must use 3d mark it should be the latest one and it should be so that you can get an idea of what future titles will run like on your card, even then take it with a scoop of salt.

You did not waste your money. SLI has its own set of headaches such as possible microstuttering, finding the right profile for each game to work effeciently, some games just never working right with it, newer games running like garbage on it, etc.

Even though your 6850 is probably only slightly faster in performance than the 6850 you have eliminated those headaches. Also it should be cooler and quieter as well. Dx11 feature set is nice to have, although you probably cannot really use tessellation - at least you will have AA in certain games like BFBC2.

 

[digitaldurandal]

Outside of tessellation, there are really not many games where DX11 by itself brings benefits (I can think of Civilization 5 and BaffleForge where the cards run faster in DX11). Most likely, we won't see "pretty" DX11 games until next generation of consoles (so 2013+).

I believe that with dx11 you can run aa even with certain lighting effects that you could not do in dx10 without a hack (ie batman, bfbc2, ue3 games mostly.)

A lot of the benefits to new dx versions are seen more on the programming side than on our side. dx versions typically become easier and easier to build a 3d game including all the latest eye candy.

 

[tyl998]

I've had no problems with SLI so far. All of my games have worked beautifully, but of course my game collection is pitiful compared to the entirety of PC games out there, so take my personal experience as you will.

 

[bryanW1995]

Outside of tessellation, there are really not many games where DX11 by itself brings benefits (I can think of Civilization 5 and BaffleForge where the cards run faster in DX11). Most likely, we won't see "pretty" DX11 games until next generation of consoles (so 2013+).

yeah, if you don't play either of those you probably won't see the benefits until future "true" dx11 games come out. however, I can say that at least in civ V it is a HUGE difference. my gtx 260 absolutely sucks in civ V, while my gtx 460 768 becomes cpu limited on all but the largest maps.

it's probably a waste of money right now but you'll probably end up getting good usage out of the card. plus it should OC very well and you won't have to worry about dual card issues so overall a +, especially if you got decent $$$ for the 9800gt's.

 

[tyl998]

Wow a game actually gets CPU limited on your i7 920? Jeez...

 

[happy medium]

yes you wasted your money, since no one else will tell you.

 

[nenforcer]

The only thing you gained was power usage and probably temperature (heat) lessened.

You should probably be able to make up the cost in your power bill after about 18-24 months.